An efficient resource model updating framework concept was proposed aiming for the improvement of raw material quality control and process efficiency in any type of mining operation. The concept integrates sensor data measured online on the production line into the resource or grade/quality control model and continuously provides locally more accurate estimates. The concept has been applied in a lignite field with the aim of identifying local impurities in a coal seam and to improve the prediction of coal quality attributes in neighbouring blocks. A significant improvement was demonstrated which led to better coal quality management. So far, the proposed concept and the application in coal mining was limited to a case where online measurements were unambiguously trackable due to a single extraction face being the point of origin for the material. This contribution presents an extension to the case, where characteristics from blended material, originating from two or three simultaneously operating extraction faces, are measured. The challenge tackled in this contribution is the updating of local coal quality estimates in different production benches based on measurements of a blended material stream. For a practical application of the updating concept, which is based on the Ensemble Kalman Filter, a simple method for generating prior ensemble members based on block geometries defined in the short-term model and the variogram, is discussed. This method allows for a fast, semi-automated and rather simple generation of prior models instead of generating a fully simulated deposit model using conditional simulation in geostatistics. It should foster operational implementation in an industrial environment. The main purpose of this article is to investigate the applicability of the developed framework with a simplified prior resource model. In addition to this any model improvements due to the integration of sensor data obtained by observing a blend of coal from multiple extraction faces is investigated.